V33F-01 INVITED 13:45h
Experimental studies bearing on the role of brines in deep-crustal metamorphism
Action of an aqueous fluid in granulite facies metamorphism can explain such commonly observed features as alkali-exchanged feldspars, garnet corrosion textures, synmetamorphic quartz veins, apparent mobility of Ca, Fe, Mg, and REE, depletion of Rb and Th, and, in some granulites, anomalously high oxidation state. Such a fluid, though geochemically effective, must have low H$_{2}$O activity to suppress hydrous minerals and prevent wholesale melting. CO$_{2}$-rich fluids have often been invoked as granulite-forming fluids, but have very low solubility for most common minerals. Concentrated chloride solutions satisfy the low H$_{2}$O activity requirement. (Na, K)Cl brines at 600-900 $\deg$C undergo pressure-induced ionization near 5 kbar, which results in a sudden decrease in H$_{2}$O activity (Aranovich and Newton, CMP 127, 261, 1997). A 30 mol % solution of (K$_{.2}$Na$_{.8}$)Cl at 8-10 kbar and 800 $\deg$C would have a(H$_{2}$O) near 0.5. Aranovich and Newton (Am. Min. 83, 193, 1998) used measured a(H$_{2}$O) of KCl solutions in equilibrium with the model granulite assemblage phlogopite-quartz-enstatite-K-feldspar to show that the threshold a(H$_{2}$O) in medium- to high-pressure granulite facies metamorphism is 0.4-0.5, much higher than previously thought. Thus, alkali chloride solutions of moderate concentration, well short of salt saturation, have suitably low a(H$_{2}$O). Recent studies in this laboratory and elsewhere reveal contrasting solubility behavior of common minerals in NaCl solutions at high T and P. Enhancement factors of molality (mol solute/kg H2O) relative to salt-free H$_{2}$O at 10 kbar, 800 $\deg$C and X(H$_{2}$O)=0.3 are 0.4 for quartz, 0.34 for albite (9 kbar: Shmulovich et al, CMP 141, 95, 2001), 16 for calcite, 120 for anhydrite, 15 for wollastonite, 34 for forsterite and 110 for fluorapatite. These data show that carbonate, sulfate, phosphate, the divalent cations, and rare earth elements are very soluble in high P-T alkali chloride solutions, and that the aluminosilicate constituents retain moderate solubility. A complex brine at high P and T could be effective in deep-crustal metasomatism, even at low fluid/rock ratios. The large solubility enhancement of anhydrite may be significant in that a CaSO$_{4}$-rich brine would necessarily be highly oxidized, with f(O$_{2}$) midway between NNO and HM in the presence of granulite facies assemblages. This could explain the high f(O$_{2}$) shown by some terranes (Wilson Lake, Labrador; Shevaroy Hills, S. India). Recent measurements (Webster et al, GCA 63, 729, 1999) have shown that Cl is very soluble in H$_{2}$O-undersaturated basalt magmas at elevated P and high T, and strongly pressure-dependent, so that Cl could cause early out-gassing during ascent of basalts, liberating concentrated brines. Thus, underplating of basalts could not only supply heat for metamorphism, but also geochemically active fluids compatible with granulite facies assemblages and only limited anatexis of country rocks.
V33F-02 14:15h
Brine Rich Diamond-Forming Fluids
Micro-inclusions in diamonds provide pristine information on the composition of mantle fluids. We explored the composition of sub-micrometer inclusions in 12 fibrous diamonds from Diavik, Slave Craton, Canada. TEM investigation of the inclusions revealed a multi-phase halide-carbonate assemblage with minor apatite. Fluid is also present indicating crystallization during cooling from a primary fluid, trapped during the diamond growth. Potassium is concentrated in halide and fluid; no K-bearing carbonates were found. Ten diamonds carry brine-rich fluid with an average composition of K$_{6}$ Na$_{4}$CaMgFeBa(Si,Al)O$_{2}$Cl$_{9.4}$(CO$_{3}$)$_{4.3}$(H$_{2}$O)$_{10}$ (determined using EPMA and FTIR). In one zoned diamond carbonatitic melt inclusions populates the rim (K$_{13}$Na$_{24}$Ca$_{11}$Mg$_{22}$ Fe$_{5}$Ba$_{2}$Si$_{7}$AlP$_{2}$Cl$_{11}$); the mantle carries brine (K$_{18}$Na$_{24}$Ca$_{6}$Mg$_{10}$Fe$_{4}$Ba$_{3}$Si$_{5}$PCl$_{26}$). One diamond carries composition intermediate between hydro-silicic and carbonatitic melt. The brine in the Canadian diamonds is similar to that found by Izraeli et al. (2001) in cloudy eclogitic and peridotitic diamonds from Koffiefontein, but is richer in Na, Fe and Ba. Micro-inclusions of peridotitic minerals were found in two of the Canadian diamonds. Integrating diamond fluid data from Africa, Brazil, Siberia, and Canada, we found a narrow ranges of fluid composition varying between four end members: hydrous melts rich in silica and alkalis, carbonatitic melts rich in Mg, Fe and Ca, brine rich in Cl, K and Na and sulfide melts rich in Fe and Ni. Carbonatitic melts were found together with all other fluids. The other three components were never detected together in any single diamond and no mixing lines were observed between them. Brine may be generated from parental carbonatitic melts by carbonate crystallization and separation of the residual melt into two immiscible fluids: brine and hydrous-silicic melt. Diamonds can grow from all these fluids. The trace element chemistry of the diamond-forming fluids is similar to that of kimberlites. It is possible that kimberlitic magmas at depth are closer in composition to the trapped fluids and to carbonate and halide-rich fluids recently found in olivine phenocrysts in an Udachnayan kimberlite. The volatile content of erupting kimberlites represents magma that degassed during most of its ascent.
V33F-03 INVITED 14:30h
Fluid Inclusion Evidence for Brines in the Earth's Crust
Over the past few decades it has become clear that brines of variable origin, salinity and chemistry are common in many crustal environments. Most of our current understanding of the distribution and chemistry of brines has come from studies of fluid inclusions trapped in minerals. Previously it was only possible to determine the bulk salinity (in terms of weight percent NaCl equivalent) and the major salts present in the inclusions. However, over the past two decades significant advances have been made in our ability to analyze individual fluid inclusions using techniques such as Raman and FTIR spectroscopies, synchrotron XRF, PIXE, PIGE and laser ablation ICP-MS. Today, it is possible to determine not only the major element chemistry but also the trace element chemistry and volatile contents of individual inclusions as small as about 10 microns. In deep sedimentary basins, brines with salinities of 30-40 wt.percent TDS (total dissolved solid) are common and are often associated with hydrocarbon reservoirs. These basinal brines usually contain Na or Ca as the dominant cation and Cl as the major anion and originate through hydration reactions and interaction with evaporites during basin evolution. Methane is a common component of these fluids. Such brines are thought to be the main source of metals to form Mississippi Valley-type Pb-Zn-Cu deposits. Brines in silicic magmatic hydrothermal systems often achieve salinities in excess of 50 wt.percent TDS. The chemistry is dominated by Na and K chlorides, and include significant amounts of Fe and other transition metals in ore-forming systems. In these magmatic systems, the high salinities are often the result of aqueous fluid immiscibility that preferentially partitions the less-volatile components into a highly saline liquid phase, while more volatile components are partitioned into a low salinity vapor. In other magmatic systems, high salinity brines are exsolved directly from the crystallizing melt - in some cases the salinities are so high that the fluids are more correctly referred to as salt melts rather than aqueous solutions. In submarine hydrothermal systems brines achieve their salinities through sub-seafloor boiling of seawater. Such brines are dominated by calcium salts and may contain significant amounts of methane. Brines are less common in metamorphic environments compared to deep sedimentary basins or magmatic systems. Metamorphic brines are thought to originate mainly as a result fluid immiscibility (as in magmatic systems) or by metamorphism of evaporates.
V33F-04 14:45h
Role of KCl-NaCl and CaCl2 Fluids in Igneous and Ore-Forming Processes
Experiments from two to nine Kilobars with KCl-NaCl and NaCl-CaCl2 fluids reveal that chloride partitions strongly in favor of the fluid phase up to six Kilobars and its partition in favor of a melt slightly increases above six Kilobars. KCl/NaCl ratio in the fluid phase in equilibrium with a granitic melt increases with increasing pressure while the KCl/NaCl ratio in the coexisting melt decreases at 700 C. KCl/NaCl ratio both in the fluid phase and the melt, however, becomes pressure independent at 750 C. At a given pressure KCl/NaCl ratio in a fluid phase coexisting in equilibrium with a granitic melt increases with increasing temperature between 650 and 750 C. This explains the K/Na zoning typically seen in granitic pegmatites. CaCl2/FeCl2 ratio in silica-rich melts and coexisting aqueous phase decreases with decreasing pressure making the fluids very rich in iron. This experimental data on the partitioning of iron at moderately low pressures supports the general conclusion that the source for iron in most contact metasomatic deposits is the intrusive and the chloride-bearing aqueous solutions contain appreciable amounts of iron, which in turn, could precipitate in reactive rocks such as carbonates.
V33F-05 15:00h
High salinity volatile phases in magmatic Ni-Cu-platinum group element deposits
The role of "deuteric" fluids (exsolved magmatic volatile phases) in the development of Ni-Cu-PGE (platinum group element) deposits in mafic-ultramafic igneous systems is poorly understood. Although considerable field evidence demonstrates unambiguously that fluids modified most large primary Ni-Cu-PGE concentrations, models which hypothesize that fluids alone were largely responsible for the economic concentration of the base and precious metals are not widely accepted. Determination of the trace element composition of magmatic volatile phases in such ore-forming systems can offer considerable insight into the origin of potentially mineralizing fluids in such igneous environments. Laser ablation ICP-MS microanalysis allows researchers to confirm the original metal budget of magmatic volatile phases and quantify the behavior of trace ore metals in the fluid phase in the absence of well-constrained theoretical or experimental predictions of ore metal solubility. In this study, we present new evidence from major deposits (Sudbury, Ontario, Canada; Stillwater Complex, Montana, U.S.A.) that compositionally distinct magmatic brines and halide melt phases were exsolved from crystallizing residual silicate melt and trapped within high-T fluid conduits now comprised of evolved rock compositions (albite-quartz graphic granite, orthoclase-quartz granophyre). Petrographic evidence demonstrates that brines and halide melts coexisted with immiscible carbonic phases at the time of entrapment (light aliphatic hydrocarbons, CO$_{2}$). Brine and halide melt inclusions are rich in Na, Fe, Mn, K, Pb, Zn, Ba, Sr, Al and Cl, and homogenize by either halite dissolution at high T ($\sim$450-700$\deg$C) or by melting of the salt phase (700-800$\deg$C). LA-ICPMS analyses of single inclusions demonstrate that high salinity volatile phases contained abundant base metals (Cu, Fe, Sn, Bi) and precious metals (Pt, Pd, Au, Ag) at the time of entrapment. Notably, precious metal concentrations in the inclusions are comparable to and often exceed the economic concentrations of the metals within the ores themselves. As a consequence of these results, current genetic models must be revised to consider the role played by hydrous saline melts and magmatic brines in deposit development, and the potential for interaction and competition between sulfide liquids (or PGE-bearing sulfide minerals) and hydrosaline volatiles for available PGE and Au in a crystallizing mafic igneous system must be critically evaluated.
V33F-06 15:15h
Chlorine cycling in subduction zones: Insights from submarine glasses and melt inclusions from arc and back-arc basalts
Chlorine is an excellent tracer of contributions of slab-derived fluids to mantle-derived melts produced during subduction zone magmatism, as saline fluids released during slab dehydration are enriched in chlorine by several orders of magnitude over mantle peridotite. Chlorine and other halogens also strongly influence the solubility of many cations in high temperature solutions and thus play an important role in mediating fluid-driven mass transfer within subduction zones. However, quantitative understanding of chlorine systematics in subduction zone magmas is limited by loss of chlorine during high level degassing. This problem may be circumvented by: (i) analyzing samples, typically from back-arc spreading systems, that erupt under sufficient depth of seawater to limit chlorine loss, and (ii) using melt inclusions hosted in early-formed phenocrysts to examine the chlorine contents of melts trapped prior to degassing. Chlorine contents of primitive back-arc basalt glasses range from values that overlap those of MORB and OIB to substantially higher values ($>$ 1000 ppm). Although chlorine contents may be strongly enriched from shallow seawater contamination processes, this is relatively easily discerned from enrichment due to slab inputs as the latter correlates strongly with other indices of slab input. By coupling measured chlorine, and H$_{2}$O contents with simple flux melting models it can be shown that the slab fluids that contribute to back-arc melts have salinities equivalent to ~1-10% NaCl, sufficient to strongly alter cation solubilities. Furthermore, the estimated Cl/H$_{2}$O and Cl/K$_{2}$O of the slab-derived components also decrease systematically with distance from the arc front, consistent with progressive preferential loss of chlorine with respect to these elements. Melt inclusions from arc lavas generally show higher chlorine and Cl/K$_{2}$O than back-arcs, consistent with greater contributions from slab fluids, but within arcs may show complex variations related to source and melting regime. For example melt inclusions from OIB-like lavas from the Cascade Arc range to higher chlorine contents and Cl/K$_{2}$O than evident in CAB and MORB-like samples, suggesting that in this hot and dry arc an enriched source region may produce greater enrichment in chlorine than inputs from slab fluids.
V33F-07 15:30h
Fluoride Melts: Occurrence, Origin and Implications for Element Transfer Processes in Subduction Zones
Hitherto unknown fluoride melts were found in metasomatized mantle xenoliths from New Zealand. These fluoride melts were only preserved in rock fragments that were carefully polished using non-hydrous polishing liquids. The protogranular spinel wehrlites consist of mm-sized olivine, clinopyroxene, amphibole, accessory minerals as apatite and spinel and, on grain boundaries and in melt pockets on triple junctions, silicate and fluoride glasses. Fluoride glasses occur as veinlets and as thin films on grain boundaries, as well as in melt pockets on triple junctions. The fluoride glass is transparent, slightly yellowish and sometimes contains small secondary clinopyroxenes, and only rarely sulfide blebs or fluid inclusions. The fluoride melts are interpreted to be derived immiscibly from a precursor silicate melt and the most spectacular textural evidence for liquid immiscibility is found in one of the xenoliths. Minerals and melts in the xenoliths were analysed for major and trace elements using electron microprobe and Laser Ablation ICPMS. Trace elements are effectively partitioned between immiscible fluoride and silicate melts. For example, separation of immiscible silicate and fluoride melts fractionates light REE from heavier REE or HFSE from REE. In many cases, silicate glasses found within mantle xenoliths are products of infiltration of the host magma into the mantle xenoliths during ascend. This is, however, not the case here as comparison of the major and trace element composition of the host lava with the silicate glass indicates. The major element composition of the immiscible silicate glasses is characterised by extreme enrichment in Mg and Al, which places them close to high-Mg magmas that are commonly found in subduction zones. The genetic link to subduction zones is further substantiated by extremely low high-field strength element concentrations (e.g., Ti, Zr, Hf) that are characteristic for magmas observed in subduction zone settings.